DiffRatio – A One-Step Diffusion Model with SOTA quality and 50% less memory(arxiv.org)
arxiv.org
DiffRatio – A One-Step Diffusion Model with SOTA quality and 50% less memory
https://www.arxiv.org/pdf/2502.08005
1 comments
This paper was posted on arXiv today. It shows surprisingly strong results on ImageNet at 512 resolution (FID 1.41) with one-step generation, while requiring 50% less training-time memory. Do you think this could become the next standard training method for image foundation models? Feel free to leave your comments.